Exon Skipping and Alternative Splicing of CPB2 mRNA in Multiple Cell Types Results in Variants of TAFI That Are Inactive and Not Secretable

Abstract 1189

Thrombin-activatable fibrinolysis inhibitor (TAFI) is a basic carboxypeptidase zymogen that plays important roles in modulation of fibrinolysis and inflammation. Activated TAFI (TAFIa) removes carboxyl-terminal lysine and/or arginine residues from substrates such as partially-degraded fibrin, cell-surface plasminogen receptors, bradykinin, the anaphylatoxins C3a and C5a, and thrombin-cleaved osteopontin. The plasma pool of TAFI arises from expression of its gene (CPB2) in the liver. However, CPB2 is expressed in other locations including platelets (arising from expression in megakaryocytes), monocytes, and macrophages. An additional source of CPB2 expression has been shown to be the hippocampus; this TAFI variant was reported to be expressed from a CPB2 mRNA in which (i) exon 7 had been skipped resulting in an in-frame loss of 37 codons and (ii) alternative splicing had occurred in exon 11 resulting in a frameshift that deletes the final 42 codons and introduces a novel 16-amino acid carboxyl-terminus. Most recently, skipping of exon 7 has been reported in HepG2 (human hepatocellular carcinoma) cells, a phenomenon that appears to play a role in balancing selection at the CPB2 locus in the human population. As much as 12.5% of the CPB2 transcript in HepG2 cells was reported to lack exon 7.

Accordingly, we have characterized, using RT-PCR, molecular cloning, and quantitative RT-PCR, the splicing patterns of CPB2 mRNA in a variety of cell types. We examined RNA isolated from human liver, HepG2 cells, the megakaryocytoid cell line Dami, platelets, the monocytoid cell line THP-1, and human cerebral cortex and cerebellum. We found evidence for alternative splicing/exon skipping in all cell types tested. All cells contained CPB2 mRNA lacking exon 7. Only platelets, cortex, and cerebellum CPB2 mRNA featured alternatively spliced exon 11, and all cDNA clones identified that contained exon 11 alternative splicing also lacked exon 7. Quantitative analysis of the proportion of total CPB2 transcripts that lack exon 7 showed that HepG2 cells had almost 10% exon 7-less transcripts but all other cell types tested had far lower proportions, ranging from 1% (Dami cells, peripheral blood mononuclear cells and cerebellum) to less than 0.1% (liver, THP-1 cells, platelets).

Studies of CPB2 expressed in the hippocampus suggested that the variant lacking exon 7 and featuring alternative splicing in exon 11 encodes a protein that is localized in the endoplasmic reticulum of neural cells and that possesses endopeptidase activity against amyloid precursor protein. To test the functional properties of the TAFI proteins encoded by the TAFI variants, we transfected baby hamster kidney cells with expression plasmids encoding variants lacking exon 7, alternatively spliced exon 11, or both variations. Interestingly, unlike wild-type recombinant TAFI in these cells, the variant proteins could not be secreted, despite the presence of an intact signal peptide in each. Western blot analyses of transfected cell lysates revealed immunoreactive bands between 40 and 45 kDa, consistent with hypoglycosylated TAFI; lysates of cells expressing wild-type TAFI contained a 45 kDa species and a 60 kDa mature preproprotein. We therefore propose that the variant proteins are aberrantly folded and thus do not exit the ER. Notably, none of the variant proteins could be activated by thrombin-thrombomodulin and they did not show activity in a specific functional assay for TAFIa.

Deletion of exon 7-encoded residues removes two surface α-helices and a single internal β-strand from the TAFI structure. Alternative splicing in exon 11 deletes a critical catalytic residue (Glu363). It is therefore not surprising that the variants are aberrantly folded, are not secretable, and lack TAFIa activity. It is also difficult to envisage how such a protein could acquire endopeptidase activity. We therefore speculate that variant TAFI resulting from exon skipping and alternative splicing may act as a chaperone for the presumptive peptidase that recognizes amyloid precursor protein. Moreover, full-length TAFI is expressed in the brain and may regulate brain-expressed tPA and plasminogen to influence neural function. Finally, it is possible that, under certain circumstances, the extent of exon skipping/alternative splicing is sufficient to impact the secretion of functional TAFI from liver or other cell types.